Click to edit Master text styles
Second level
Third level
Fourth level
Fifth level
‹date/time›
‹footer›
‹#›
------=_NextPart_01C53EDA.3AC63CB0
Content-Location: file:///C:/ED3412B3/csi-2005-Spears_files/master03.xml
Content-Transfer-Encoding: quoted-printable
Content-Type: text/xml; charset="utf-8"
CSI-Miami
1
A broad SIDE: The cognitive and strategic <=
/span>=
effects of computer-mediated communication
•Russell Spears
–Cardiff University/ University of Amsterdam
•Martin Lea
–University of Manchester
•Tom Postmes
–University of Exeter
•Anka Wolbert
–University of Amsterdam
•Sue Watt
–University of New England
–
CSI-Miami
2
A broad SIDE: The cognitive and strategic <=
/span>=
effects of computer-mediated communication
Structure of talk
• SIDE overview:
• Background and a short history (BackSIDE)
=
• Some old/borrowed studies
• New research on gender:
<=
span
style=3D'font-family:Arial;font-size:71%'>•Theory: cognitive vs strategic
SIDE
<=
span
style=3D'font-family:Arial;font-size:71%'>•Evidence: Classic cognitive effe=
cts:
<=
span
style=3D'font-family:Arial;font-size:64%;visibility:hidden'>–<=
/span>=
(Getting in touch with our male a=
nd
female SIDE)
<=
span
style=3D'font-family:Arial;font-size:86%'>–Some new cognitive effects
<=
span
style=3D'font-family:Arial;font-size:86%'>–Some new strategic effects
3
The Social-Identity model of
Deindividuation Effects (SIDE): Background
Theoretical resources
•Social identity/self-categorizat=
ion
theory
•Depersonalization when immersed =
in
groups
–Depersonalization is not
deindividuation!!
•Anonymity can depersonalize bein=
g,
behavior, social perception
Background/history
•Reicher’s work on crowd be=
havior
and critique of deindividuation theory
–No loss of identity when immerse=
d in
the crowd
–Switch from personal to social
identity!
–Could the same apply in anonymou=
s CMC?
(Yes!)
•I.e. more group salience, more g=
roup
influence in CMC?
4
![](3D"slide0032_image001.gif")
The Social-Identity model of
De-individuation Effects (the cognitive SIDE)<=
/span>
Group
identity
Personal
=
;
identity
Anonymity
Increased
group
salience
Reduced
group
salience
Adherence
to
group
norms
Adherence
to
Personal
norms
=
Salient
identity
=
Contextual
factors
=
Outcome
•=
P=
rocess
5
![](3D"slide0033_image002.gif")
The Social-Identity model of
De-individuation Effects (the cognitive SIDE)<=
/span>
Group
identity
Personal
=
;
identity
Anonymity
Increased
group
salience
Reduced
group
salience
Adherence
to
group
norms
Adherence
to
Personal
norms
=
Salient
identity
=
Contextual
factors
=
Outcome
•=
P=
rocess
![](3D"slide0033_image003.gif")
=
Group
=
Polarization
6
There’s more!: A strategic side of anonymity
= (and SIDE)
= (and SIDE)
Sometimes we can hide behind ano=
nymity
to escape =
the
consequences of power…
•There is a strategic side to SIDE!
–Spears & Lea (1994)
–Reicher, Spears & Postmes (1995)
•We can ignore/challenge powerful outgroups when:
–We are anonymous to them (CMC)
–Or have strength in numbers (solidarity).
•Other strategic effects?
–Identity dissimulation, e.g., gender-bending
7
The cognitive SIDE:
Evidence?
(review old studies: backSIDE)
(review old studies: backSIDE)
•Evidence of greater group polari=
zation
in a=
nonymous
CMC groups
–Spears Lea, & Lee, 1990
=
–<=
span
style=3D'font-family:Arial;font-size:86%'>…but only when group ident=
ity
(not individual identity) salient.
–This is implicit evidence for
conformity to group norms…
8
Spears, Lea & Lee (1990):
Group polariz= ation in CMC
Group polariz= ation in CMC
Salient
identity
9
More evidence?
•But can we find more explicit ev=
idence
for conformity to group norms under conditions of anonymity in CMC?
=
=
10
Postmes, Spears, Sakhel & de Groot (2001)
Design
• 2 =
Norm
=
–(Prosocial
vs Efficiency-oriente=
d)
•x2=
Anonymity
<=
span
style=3D'font-family:Arial;font-size:120%'>–(Anonymous
vs visible)
•<=
span
style=3D'font-family:Arial;font-size:117%;mso-special-format:lastCR;displa=
y:
none'>
11
Manipulation of Social Norm:<=
/span>
•Scrambled sentence priming proce=
dure
–Efficiency-oriented norm
–e.g. “The efficiently work=
is in
office the remarkably done” >
–“The work in the office is=
done
remarkably efficiently”
–<=
span
style=3D'font-family:Arial;font-size:86%;display:none'>
=
–Prosocial norm
–e.g. “The way support other
group each the all” >
–“The group support each ot=
her
all the way”
12
Manipulation of anonymity:
•A digital photo was made of all
participants who are then placed in computer cubicles
•=
•Visible condition:
<=
/span>
–Photos are displayed during group
discussion
•Anonymous condition:=
–Photos are not displayed<=
span
style=3D'font-family:Geneva;font-size:86%;mso-special-format:lastCR;displa=
y:
none'>
13
Procedure
1.Scrambled
sentence task (individually)
2.Group
task…
–<=
/span>to create group identity
3.Group
discussion via CMC
14
Group discussion scenario:
•Problems have arisen in a medium=
sized
hospital. Through an increase in the number of patients the hospital has problem=
s providing adequate care to all
patients in time. The management of the hospital wants to solve the problem by increas=
ing
the =
efficiency
in the organization.... The doctors and nurses want to secure patient welfare, and prev=
ent
patients being treated as numbers.... Discuss possible solutions to this problem
with your g=
roup.
15
Dependent measures:
•Solution to dilemma (open-ended)=
–Coded for efficiency/prosocial
orientation
•Word counts
–Efficiency vs. prosocial words
used
–(e.g., task, efficient, costs vs.
feeling, personal, treatment)
16
Solutions proposed after
discussion:
Study 1.
Study 1.
17
Conclusions
•Greater norm-directed social inf=
luence
under =
conditions
of anonymity than visibility
• Support for SIDE
• But is this true GROUP influence...
• ….or just individual priming?
=
18
Study 2: Rationale
&#=
8226;If we prime just half the group =
and
find that i=
nfluence
generalizes to the other half, then this is true social influence!
19
Study 2: Design
•
• =
2 Anonymity (Anonymous vs Identifiable) Between Groups=
•x 2 Prime (Task-oriented
vs Neutral) =
Within
Groups
•
•Groups of 4 (2 primed with task-orientation, 2=
neutral)
20
Word Use: Efficiency-oriented
Words
21
Conclusions Study 2
•Influence does transfer to non-primed group members
•Influence in the non-primed members increases with exposure to the primed others=
3;
•Some evidence of mediation by group salience=
•
•
•So there is influence, and it is social,
•and group-based!
•
22
Study 3: Further process evid=
ence
and SIDE effects on group attraction
Lea, Spears & De Groot (2001=
)
•Theory
–Following SIDE, SCT variables sh=
ould
mediate between anonymity and group attraction
•Design
–Manipulate anonymity in group
discussions
=
–<=
span
style=3D'font-family:Arial;font-size:71%'>Measure mediating SCT variables
(self-categorization, self-stereotyping)
–Assess impact on group attractio=
n
23
Lea, Spears & De Groot (2001):
SIDE effects=
SIDE effects=
24
Conclusions Study 3…
•Anonymity effects mediated by SCT variables (depersonalization process)
=
;
•Anonymity in the group can strengthen group attraction, a group-level social
effect
•
•
<=
/span>
25
But does anonymity always mean
more group effects? (New studies)
•No!
(SIDE spin?!)
1.Depends
on identity salience
–Spears, Lea & Lee (1990)
•
26
Spears, Lea & Lee (1990):
Group polariz= ation in CMC
Group polariz= ation in CMC
Salient
identity
27
But does anonymity always mean
more group effects? (New studies)
No! (SIDE “spin”?)&#=
13;
<=
span
style=3D'font-family:Arial;font-size:71%'>–Spears, Lea & Lee (1990)=
<=
span
style=3D'font-family:Arial;font-size:71%'>–Tom: next talk (inductive vs. deductive)=
<=
span
style=3D'font-family:Arial;font-size:75%'>3.(This talk) Depends on key features of the group
–Some groups/categories do become more salient through=
anonymity-depersonalization (the
cognitive SIDE so far…)
<=
span
style=3D'font-family:Arial;font-size:83%'>•E.g., students, nationality
–But some groups may become cued, more salient, throug=
h visibility=
<=
span
style=3D'font-family:Arial;font-size:83%'>•E.g. gender, ethnicity
28
Lea, Spears & Watt (2005)=
: Vi=
sibility
and anonymity effects on attraction and gro=
up
cohesiveness
•Rationale/predictions:
•For
non-visually cued categories…
–E.g. nationality
•Anonymity fosters self-cat, group attraction, etc (classic SIDE)
=
–(”inside” route)
•For
visually cued categories…
–E.g. gender
•Visual exposure cues fosters self-cat, group attraction etc
–(”outside” route)
•For
all categories…
–Visibility cues interpersonal attraction (wait for Tom…)
•Group
attraction stronger under high category salience
•Interpersonal
attraction stronger under low category salience
29
Lea, Spears & Watt (2005)=
: Vi=
sibility
and anonymity effects on attraction and gro=
up
cohesiveness
•Design
details (2x2x2)
•36
4-person CMC groups discussing (2+2=3D4) topics
•Group identities
–Nationality: 2 Brits vs. 2 Dutchies (separate countries)=
–Gender: 2 males vs. 2 females (separate rooms)
------=_NextPart_01C53EDA.3AC63CB0
Content-Location: file:///C:/ED3412B3/csi-2005-Spears_files/slide0084.htm
Content-Transfer-Encoding: quoted-printable
Content-Type: text/html; charset="us-ascii"
•Visibility vs anonymity
–Live streaming web-cam video vs not.
•Group salience: discussion topic (within Ps)
–High salience (2 topics: G vs. N) relevant to salient category
–Low salience: 2 topics relevant to other category
<=
/div>
•E.g. nationality measures after talking about gender topic
•Topic order counterbalanced
•Measures after each set of (2) topics
30
Lea, Spears & Watt (2005)=
: Vi=
sibility
and anonymity effects on attraction and gro=
up
cohesiveness
•Constructs/measures
(all 2 items per scale)
•Self-stereotyping
–Nationality: ‘participant n(=3Dself) is typically
British/Dutch’
–Gender: ‘participant n(=3Dself) is typically male/female’=
=
•Self-categorization
–Nationality: e.g. ‘During the discussion I felt
British/Dutch’
–Gender: e.g. ‘During the discussion I felt male/female’&#=
13;
•Attraction to own category
–Nationality: ‘I liked the British/Dutch… ‘
–Gender: ‘I liked the males/females’
•Cohesiveness
–Nationality: e.g. ‘…there was a feeling of unity among
British Ps’
–Gender: e.g. ‘…there was a feeling of unity among the fem=
ale
Ps’
31
Fi=
g. 1.
Nationality category: anonymous depersonalized attraction (high salience)
and visually-cued interperson= al attraction (low salience).
and visually-cued interperson= al attraction (low salience).
![](3D"slide0075_image020.gif")
32
Fi=
g. 2.
Gender category: visually-cued depersonalized attraction (high salience)
and visually-cued interperson= al attraction (low salience).
and visually-cued interperson= al attraction (low salience).
![](3D"slide0076_image023.gif")
33
Lea, Spears & Watt (2005)=
: Vi=
sibility
and anonymity effects on attraction and gro=
up
cohesiveness
•Summary/Conclusions
•Anonymity can enhance group effects in certain =
groups/categories
–Where category is not visibly cued
•Visibility can enhance group effects in other groups/categories
<=
/span>
–Where social category is visibly cued
•Visibility enhances interpersonal processes
–Tom will elaborate this theme!
–
•So was classic cognitive SIDE wrong?
–No! Just partial, provisional, progressive
–Contradictions? What contractions?=3DSIDE spin!
34
But what about the strategic side of anonymity? (and SIDE?)
Sometimes we can hide behind ano=
nymity
to escape =
the
consequences of power…
•There is a strategic side to SIDE!
–George Bush: “You can run but you can’t hide”
<=
/span>
–In cyberspace you can both hide
and stay put!
–“CMC is on your SIDE!”
•We can ignore powerful outgroups when:
–We are anonymous to them
–Or have strength in numbers (solidarity).
35
But what about the strategic side of anonymity? (and SIDE?)
CMC provides
–Anonymity (unless surveillance!)
–Communication channel for support
–i.e. medium of resistance to power!
Some evidence:
Spears, Lea, Corneliussen Postmes & Ter Haar (200=
2)
–greater conformity to ingroup norms punishable by outgroup when group has CMC
available
–i.e. better able to communicate support and coordinate resistance with CMC t=
han
when isolated.
36
But what about the strategic side of anonymity? (and SIDE?)
•Can we use (anonymous) CMC to avoid a <=
span
style=3D'position:absolute;top:34.0%;left:19.85%;width:64.41%;height:5.25%=
'>low status identity altogether?&=
#13;
–Dissimulation
–e.g. gender-bending…
–
37
Theoretical background…
•Some have argued that cmc/intern=
et reduces power/status differences…&=
#13;
–Because of anonymity (reducing
accountability)
–e.g., Kiesler & colleagues&#=
13;
•…by allowing freedom to
choose/shape own identities
–Because of anonymity
=
–e.g. Turkle, Herring, Harraway, =
etc=
38
Theoretical background…
•Others have argued that cmc/internet can =
increase power/status
differences…
–<=
/span>Because of anonymity
–<=
/span>(e.g., Spears, Lea, Postmes & colleagues)&=
#13;
•…by increasing:
–<=
/span>the salience of social identities
=
•<=
/span>(the cognitive side of SIDE)
–<=
/span>surveillance
•<=
/span>(the strategic side of SIDE)
39
Theoretical background…
•So, two sides to Internet and power…
–Positive/optimistic
•<=
/span>Liberating
–Negative/pessimistic
•<=
/span>Reinforces power
•How does this affect gender?
•<=
/span>Gender has power/status dimension
<=
/span>
•<=
/span>Internet is a relatively male domain/subcultur=
e
40
Girls will be boys: Gender identity management on the Internet
•Anka Wolbert=
–University of Amsterdam
41
Gender-bending as a strategy of resistance to (male) power
•Do women use the gender-bending potential in CMC and on the Internet more than me=
n?
•because of lower power/status?
•to participate, to be taken seriously?
•Key questions:
•Do women shift to male (high status) identities when given the choice?
•Do they only choose gender identity to suit the topic of discussion?
•Do both females and males shift identity to match topic?
42
Design…
=
43
Procedure…
•
•Ps to enter a chat room to discuss one of 5 topics
–Topics piloted for male vs. female gender domain
•Ps asked to choose “avatar” to represent them in <=
/span>chat room for each topic.
–Avatars piloted for male, vs. neutral vs. female gender=
•Discussion never actually takes place…
–choice of avatar is dependent variable
44
Avatars used in research…
•Avatars
were piloted and
grouped in 3 cl=
asses:
–1 =3D
male
–2 =3D neutral
–3 =3D female
![](3D"slide0068_image026.gif")